BF: Added out_file to SPM ApplyTransform

[swederik]

I just tried to use this interface and it didn't seem to apply the manually estimated reorientation matrix correctly.

Took a look at the code and it seemed to be applying an inverted version of the matrix. So I fixed that and just copy/pasted the code from spm_run_reorient.m. Now it will also create a new file with the applied transformation.

Note that you won't be able to actually SEE the proper reorientation until the file is resliced (even to the same resolution it is current at..). I am unsure whether or not we should include that step explicitly in this interface. My current workaround is to use Freesurfer's mri_convert to conform the reoriented file.

Seems better to have it create a new file than dangerously modifying the input file.

[mwaskom]

Option 2: ban SPM.

[swederik]

Yeah wish we could...
You know a better program for manual coregistration? A colleague was using MINC but apparently it changed the datatype of the images upon saving, leading to wildly incorrect values.

[inati]

AFNI

[mwaskom]

If you're using Freesurfer already, I've never had bbregister fail me

[swederik]

Has to be manual, I think. I really only need it in cases of severely brain damaged patients. The manual registration is mainly for putting the head in a reasonable orientation because they can't be positioned properly in the scanner.

[ag-bell]

Are you sure you need the head in a "reasonable" orientation for registration to work? I have gotten
bbregister to work on only partial brain acquisitions of fmri to T1. That is part of its charm. I am not sure what exactly you are aligning to what though.

[satra]

freesurfer bbregister would requires a surface, which might be difficult in case of brain damage. however, you might be able to get by with flirt bbr, which implements volume based bbregister.